Gulf of Mexico Science, 2006(1/2), pp. 11–13

Cannibalism of Juveniles by Adult Common Snook ( undecimalis)

A. J. ADAMS AND R. K. WOLFE

Cannibalism is probably common to many fish , but has been docu- mented for relatively few. This study provides the first documentation of inter- cohort cannibalism (juveniles cannibalized by adults) for common snook, Centro- pomus undecimalis, all instances of which occurred in putative nursery habitats.

annibalism is probably common in many Adult snook (Ͼ300 mm SL) enter Charlotte C fish species (FitzGerald and Whoriskey, Harbor mangrove creek habitats sporadically 1992), but has been documented for relatively during summer, but use these habitats fre- few. For example, Smith and Reay (1991) quently during colder winter months (A. J. Ad- found documentation of cannibalism in 36 of ams and R. K. Wolfe, unpubl. data) when age- 410 teleost families, but conjectured that can- 0 juveniles are most abundant, so are potential nibalism has likely been observed in many ad- predators of juvenile snook. Because juvenile ditional species. If a common occurrence, can- and adult snook occur concurrently in winter, nibalism can have population regulation im- a study was conducted to determine whether plications (reviewed in Smith and Reay, 1991). cannibalism occurred in these shallow creek Cannibalism of younger (smaller) individuals and tidally connected pond habitats. It is also by older (larger) conspecifics (intercohort can- possible that intercohort cannibalism can oc- nibalism) is the most common form of canni- cur between juvenile year classes (i.e., age-1 balism in fishes (Smith and Reay, 1991). For cannibalize age-0; age-2 cannibalize age-1). species with potential for intercohort overlap This paper is the first to document intercohort in habitat use, cannibalism rates may have im- cannibalism by wild C. undecimalis. [Cannibal- portant management implications (e.g., Mann, ism by C. undecimalis in aquaculture settings is 1982). common (N. Brennan, pers. comm.); however, Common snook, Centropomus undecimalis aquaculture-based cannibalism by fishes is (Bloch), is a tropical and subtropical, estua- common, even for species with little or no can- rine-dependent, species that is eco- nibalism in the wild (reviewed in Smith and logically and economically important through- Reay, 1991).] out its range, especially in Florida (Taylor et al., 2000). The general pattern of its life his- MATERIALS AND METHODS tory is known, but ecological details are lacking for numerous life stages. The general life his- Charlotte Harbor is a 700-km2 subtropical, tory is as follows: adult snook spawn in passes coastal plain that receives freshwater and inlets at the mouths of (Taylor et from three rivers and numerous mangrove al., 1998); larval planktonic stage is approxi- creeks, and is connected to the Gulf of Mexico mately 2 wk (Peters et al., 1998); juvenile hab- through five inlets (Hammett, 1990; Poulakis itats are shallow, complex, meso- to oligohaline et al., 2003). Sampling of adults for incidence habitats (Peters et al., 1998; A. J. Adams and of cannibalism occurred in four mangrove- R. K. Wolfe, unpubl. data). This life history lined creeks and four tidally connected ponds strategy serves to decouple local reproduction on the eastern shoreline of the estuary in two from recruitment of new individuals into the winter seasons: year 1, Jan.–April 2003; year 2, population. In Charlotte Harbor, Florida, the Nov.–Dec. 2003, Feb.–March 2004. Snook location of this study, postsettlement [Ͻ 20 Ն140 mm SL were captured using a 30 m ϫ mm standard length (SL)] and juvenile snook 1.8 m center bag seine (25-mm mesh). Snook occupy shallow creeks and tidally connected were measured (SL), and stomach contents ob- ponds fringed by red mangroves (Rhizophora tained at the site of capture by lavage: a semi- mangle) (A. J. Adams and R. K. Wolfe, unpubl. flexible plastic tube was inserted down the data). Juvenile snook reach 150–180 mm SL by throat into the stomach, and ambient water 1 yr of age and 300 mm SL by 2 yr (Taylor et was pumped into the stomach to force out the al., 2000), and in Charlotte Harbor remain stomach contents. Once lavage was complete, abundant in or near these habitats throughout snook were released. Effluent containing stom- the year until reaching 250–300 mm SL. ach contents was examined on site to deter- 12 GULF OF MEXICO SCIENCE, 2006, VOL. 24(1/2)

TABLE 1. Standard length (mm) of cannibalistic specific (Dominey and Blumer, common snook, Centropomus undecimalis and of in- 1984). However, in this study, the simple mix- dividuals they cannibalized. ing of size classes was not sufficient to cause cannibalism—there was no intercohort canni- Cannibalistic snook Prey snook balism observed in smaller fish. 278 48 For some species, cannibalism becomes 313 77 more common as prey becomes scarce or as 312 113 prey quality declines (reviewed in Smith and 591 117 Reay, 1991). In this study, the 10 most abun- 610 121 dant prey species (e.g., Lucania parva, Menidia 651 141 peninsulae) were small (Ͻ50 mm SL), and were likely a suitable prey base for young juvenile snook. In contrast, juvenile snook (ranked mine whether juvenile snook were present. 10th in total abundance of all creek fishes) Only instances in which juvenile snook were represented the largest individuals of the 10 readily recognized in the field were recorded most abundant species, so may have represent- as cannibalism (i.e., stomach content evalua- ed a higher quality prey source for larger tion was cursory, so did not include examina- adults. Thus, cannibalizing juveniles may have tion of otoliths, scales, or vertebrae), thus mak- been nutritionally more efficient for adult ing this an extremely conservative estimate. snook if overall prey quality was low (reviewed in FitzGerald and Whoriskey, 1992). RESULTS AND DISCUSSION Diet analysis of adult snook in open estua- rine habitats of Charlotte Harbor further sug- Stomach contents of 127 and 153 snook gests prey size preference by adults that might were examined in year 1 and year 2, respec- increase the chance of cannibalism in man- tively. Three instances of cannibalism occurred grove creek and tidal pond habitats. Although in each year. All instances of cannibalism were there was no cannibalism in 694 adult snook recorded in larger snook (smallest cannibalis- examined in open estuarine habitats of Char- ϭ tic snook 278 mm SL; Table 1), even though lotte Harbor, prey size (diet composed primar- size range of lavaged snook was 141–660 mm ily of fish) and snook size were significantly ϭ SL (mean 278.6). Mean size of cannibalistic and positively related (Blewett et al., 2006), snook was greater than noncannibalistic snook and this was the trend in this study (Table 1). ϭ Ͻ ϭ (t 1.969, P 0.05, df 277). In Year 1, Other species in which cannibalism occurs also 65.8% of captured snook had food in their tended to select for larger prey ( Juanes, 2003). stomachs, and of those with food, cannibalism In addition, the percentage of snook stomachs occurred in 3.8%. In Year 2, 58.6% of snook containing food was similar between Blewett et had food in their stomachs, and 3.4% of these al. (62%) and this study. Thus, it appears that fish exhibited cannibalism. Because the esti- the co-occurrence of adult and juvenile snook mate of cannibalism in this study is extremely in putative nursery habitats (mangrove creeks conservative (lavage may not have completely and tidal ponds) during winter exposes the ju- sampled stomach contents; stomach content veniles to potential conspecific predators. examination was cursory—only easily recogniz- Cannibalism of juveniles by adult snook was able remains were classified as snook), the rate observed in successive years, there was a posi- of cannibalism may be higher. Size of canni- tive relationship between cannibalistic snook balistic adults and prey juveniles was positively and prey snook size, and adult and juvenile correlated (Pearson correlation coefficient ϭ snook seasonally co-occur in putative nursery 0.81, P Ͻ 0.05; Table 1). All cannibalized ju- habitats. Given that a theorized attribute of venile snook were Ͻ age 1. nursery habitats is lowered predation risk, doc- Cannibalism is most common among pisciv- umented sources of predation in nursery hab- orous fishes (reviewed in Smith and Reay, itats should be investigated. 1991), and has been attributed to numerous factors. In general, cannibalism occurs when larger fish prey upon smaller fish—a relation- ACKNOWLEDGMENTS ship common to typical predator–prey rela- tionships. Thus, cannibalism may be similar to This research was funded by a grant from a predator–prey relationship when disparate Mote Scientific Foundation (AJA). We thank size classes of a species co-occur (e.g., Buckley M. Newton, K. Laakkonen, and P. Stevens for and Livingston, 1997), so may simply be intra- field assistance, and K. Leber and two anony- ADAMS AND WOLFE—CANNIBALISM IN CENTROPOMUS UNDECIMALIS 13 mous reviewers for helpful comments on the U.S. Geological Survey Water Supply Paper 2359 A, manuscript. St. Petersburg, FL. JUANES, F. 2003. The allometry of cannibalism in pi- scivorous fishes. Can. J. . Aquat. Sci. 60:594– REFERENCES 602. MANN, R. H. K. 1982. The annual food consumption and prey preferences of pike (Esox lucius)inthe BLEWETT,D.A.,R.A.HENSLEY, AND P. W. STEVENS. River Frome, Dorest. J. Anim. Ecol. 51:81–95. 2006. Feeding habits of common snook, Centropo- PETERS, K. M., R. E. MATHESON JR., AND R. G. TAYLOR. mus undecimalis, in Charlotte Harbor, Florida. Gulf 1998. Reproduction and early life history of com- Carib. Res. 18:1–14. mon snook, Centropomus undecimalis (Bloch), in BUCKLEY,T.W.,AND P. A. L IVINGSTON. 1997. Geo- Florida. Bull. Mar. Sci. 62:509–529. graphic variation in the diet of Pacific hake, with POULAKIS,G.R.,D.A.BLEWETT, AND M. E. MITCHELL. a note on cannibalism. California Cooperative 2003. The effects of season and proximity to fring- Oceanic Fisheries Investigations. Report 38:53–62. ing mangroves on -associated fish com- DOMINEY,W.J.,AND L. S. BLUMER. 1984. Cannibalism munities in Charlotte Harbor, Florida. Gulf Mex. of early life stages of fishes, p. 43–64. In: Infanti- Sci. 21:171–184. cide. Comparative and evolutionary perspectives. SMITH,C.,AND P. R EAY. 1991. Cannibalism in teleost G. Hausfater and S. B. Hrdy (eds.). Aldine, New fish. Rev. Fish Biol. Fish. 1:41–64. York. TAYLOR, R. G., J. H. GRIER, AND J. A. WHITTINGTON. FITZGERALD,G.J.,AND F. G. WHORISKEY. 1992. Em- 1998. Spawning rhythms of common snook in pirical studies of cannibalism in fish, p. 239–255. Florida. Fish. Soc. Brit. Is. 53:502–520. TAYLOR, R. G., J. A. WHITTINGTON,J.H.GRIER, AND In: Cannibalism: ecology and evolution among di- R. E. CRABTREE. 2000. Age, growth, maturation, verse taxa. M. A. Elgar and B. J. Crespi (eds.). Ox- and protandric sex reversal in common snook, ford University Press, Oxford, England. Centropomus undecimalis, from the east and west HAMMET, K. M. 1990. Land use, water use, stream- coasts of Florida. Fish. Bull. 98:612–624. flow characteristics, and water quality characteris- tics of the Charlotte Harbor inflow area, Florida. Date accepted: December 21, 2005.